Throughout this application, various references are cited in square brackets to describe more fully the state of the art to which this invention pertains. The disclosure of these references is hereby incorporated by reference into the present disclosure.
Identification of pathological or abnormal tissues has crucial importance during the diagnosis and treatment of malignant processes. Generally, cancer is diagnosed based on information obtained by using imaging methods. Certain imaging methods (CT, MRI) do not provide sufficient information for the identification of malignant proliferations, but give high resolution imaging. Other methods, especially nuclear imaging techniques provide relatively poor resolution, however, easily identify proliferating tissue parts, including various types of cancer. Thus, combinations of two types of imaging methods (PET/CT, PET/MRI) are used for the identification and proper localization of cancer.
Accurate diagnosis is generally obtained by histology or cytology. Histology is the gold standard method for abnormal/pathological tissue identification, hence tissue classification is based on histological examination of tissue specimens. Histology traditionally involves the following steps: (1) sampling (biopsy or surgery), (2) Fixation of sample using mainly formalin, (3) Processing or embedding sample into solid matrix, (4) sectioning to obtain 2-10 μm thick sections, (5) staining and (6) visual examination of sections under microscope. Staining fundamentally determines the type of information obtained. Traditional stains (e.g. eosin-hematoxylin) enable identification of cells based on morphological features, while immunohistochemical staining reveals the presence of certain proteins in cells.
As an alternative to histology/histopathology, cytopathological methods are also widely used. In case of cytopathology, only cells are taken as sample, from either biological fluids or directly from bulk tissue (aspiration cytopathology) and samples, similarly to histology, are examined under microscope after proper staining procedure.
Both histopathology/cytopathology and imaging methods are successfully used for diagnosis of cancer and follow-up of anti-cancer therapy. However, in contrast to the amount of information available before and after surgery, there is only little information available for the surgeon about the actual position of malignant tissue relative to visible features on surgical site. In the general case, surgeon relies on pre-operative imaging and his/her own senses, with special regard to tactility and vision.
The problem of positioning the malignant tissue has been traditionally solved by intra-operative histopathological examination of removed tumour. This is performed by freezing the freshly removed tissue, and sending it to pathology lab, where sample is sectioned, stained and examined under microscope. The aim of the procedure is to find out whether all the borders of removed tissue are “clear” (i.e. only healthy tissue was dissected) or not. Although the procedure is used widely, it has number of disadvantages, including about 20 minutes time demand, while the patient is in the operating room with open an surgical wound, and low reliability of the results caused by sub-optimal processing of samples.
Further methods developed for intra-operative localization of tumours include the utilization of various imaging methods during surgery. Sonography and X-ray fluoroscopy has long been used to follow surgical procedure, though their application generally causes interruption of the surgical intervention. Recently, special imaging systems based on MRI and CT have been developed for providing real-time information for surgeons. Intra-operative imaging systems have recently been equipped with navigation, which helps to link images to visually observable features. Although these systems were proven to be extremely useful in certain applications, e.g. spinal surgeries, they are not capable of identifying minor amounts of tumour tissue on surgical area or minor proximal metastases.
A promising group of recently developed technologies employs selective chemical labelling of malignant tissue. Labelling molecules carry either radionuclides or fluorescent moieties. Since proliferating cells accumulate these molecules, they can be visualized either with a gamma camera or an infrared camera for example. These methods are successfully used for detection of proximal metastases, e.g. detection of so-called sentinel lymph nodes that accumulate tumour cells close to primary tumours. Weakness of these methods lies in their selective nature to certain tumours, their incompatibility with surgical techniques and the undesired side effects of labels. It has to be noted, that melanoma can be detected by means of near-infrared two photon laser induced fluorescence without labeling, however this technique can only be used for detection of primary melanoma on skin surface.
Malignant tumours can be generally differentiated from healthy tissues based on their accelerated metabolism. Tumour cells accumulate basic nutrients or molecules that are similar to these basis nutrients (e.g. fluorodeoxyglucose—FDG). When these nutrients or fake nutrient molecules are labelled with radionuclide (I 8FDG in PET) or fluorescent moiety, tumour becomes visible using appropriate visualization method. Besides accelerated metabolism, tumours are different from healthy cells in a number of different ways. Tumours, for example show markedly different chemical composition from the distribution of small metabolic constituents to different protein expression and posttranslational modification patterns. These chemical features can be used in immunohistochemical visualization of tumours, and also in chemical imaging of tissue sections using infrared spectrophotometry or mass spectrometry. Among these methods, mass spectrometry is the sole technique which can be the basis of an in-situ, in-vivo tissue identification tool utilizing the different chemical composition of different tissues.
Mass spectrometric ionization methods have been traditionally developed for the analysis of gaseous or volatile materials. One disadvantage of these ionization methods is that they lack the capability of analysis of non-volatile compounds. This group of compounds includes peptides, proteins, nucleic acids, carbohydrates, etc.; that is approximately 90% of biologically relevant molecules.
From the 1970's, a new family of ionization methods has been developed, which was able to convert condensed phase molecules directly into ions on the gas/solid or gas/liquid interface, and subsequently desorb the nascent ions from the surface. These ionization methods are generally termed as ‘desorption ionization’ methods.
Second generation of desorption ionization methods employed an alternative way of ionization by utilizing a so-called analytical beam for ionization. Analytical beam comprises high energy particles (atoms, molecules, atomic or molecular ions, photons, etc.) which are directed onto the surface of the sample. Impact of the analytical beam on the surface produces micro-explosions yielding gaseous ions and molecules of surface material. An early method utilizing analytical beam was plasma desorption ionization which employed high energy particles produced by radioactive decay of californium isotopes [Macfarlane R D, et al. Science, 191 (4230), 920-925. 1976].
While plasma desorption utilized a divergent beam of poorly defined species, secondary ion mass spectrometry (SIMS) employed a collimated beam of atomic or cluster ions accelerated by static electric fields into the range of I 0-30 keY [Benninghoven, A, Surface Science 28(2) 541-1971]. SIMS is capable of reaching as good as I 0 nm spatial resolution, due to the cross section of focused ion beams. In spite of the excellent spatial resolution, widespread application of SIMS is strongly hindered by limited molecular weight range of molecules, which undergo SIMS ionization, Generally molecules having molecular weight below I kDa can be detected by means of SIMS, however there is a strong discrimination against heavier ions even in this narrow mass range. Method can also be used for in-depth analysis (dynamic SIMS) however in this case the higher energy ion beam produces mainly atomic ions. Investigation of liquid samples has also been developed in the case of SIMS ionization. (liquid SIMS; LSIMS) [Aberth, W, Analytical Chemistry, 54 (12): 2029-2034 1982]. Liquid-SIMS has numerous advantages compared to the original technique, including wider mass range (MW<IO kDa) Better reproducibility and sensitivity. One disadvantage of LSIMS is that samples have to be dissolved in glycerol or nitrobenzyl-alcohol prior to analysis. This step often involves solubility problems, and dissolution of solid samples obviously excludes any kind of spatially resolved analysis. Further disadvantages include the milder, but still existing limitation on molecular weight of species ionized this way.
The LSIMS method was further developed by substituting the primary ion beam with a beam of high velocity noble gas atoms. This latter technique was termed ‘fast atom bombardment’ (FAB) and had incremental advantages compared to LSIMS [Williams, D H et al, JACS, 103 (19): 5700-5704 1981], however the method kept practically all disadvantages of the original method, including strong limitations on molecular weight and loss of capability of spatially resolved analysis.
Another direction of development of the SIMS technique was to increase the mass of projectile (primary) ions. Eventually this research has led to the development of so-called massive cluster impact (MCI) ionization which utilizes multiply charged liquid (usually glycerol) droplets as projectiles in a SIMS-like experimental setup [Mahoney, J F Rapid Communications in Mass Spectrometry, 5 (10): 441-445 1991]. Droplets are accelerated to 2-10 keV/charge and high energy droplet beam is directed onto surface carrying sample material, which can be both in solid or liquid form. Substantial advantage of MCI compared to SIMS is the further extended mass/charge range and even more importantly the fact that MCI produces predominantly multiply charged ions of macromolecular species such as proteins. This advantage allows obtaining detailed mass spectrometric information, for example sequencing of proteins. MCI still carried the disadvantage of limited molecular weight range, complicated instrumentation and cross contamination between samples due to sputtering effect of impacting glycerol droplets. Although the method is theoretically capable of spatially resolved analysis, known prior art attempts to develop this capability have all failed.
A common disadvantage of the described methods is that they generally work strictly under high vacuum conditions. Hence, samples are introduced into the high vacuum regime of mass spectrometers, which involves strong restrictions on the composition and geometry of samples, and also requires special sample introduction systems.
Laser desorption ionization methods have been developed from the early 1980's [Cooks, R G et al. JACS, 103 (5): 1295-1297, 1981]. Simple laser desorption ionization, similarly to SIMS, gives poor ionization efficiencies and they can only be used for the investigation of a relatively limited number of molecules. Laser desorption methods were revolutionized by the application of so-called matrix compounds. Matrix compounds are generally mixed to samples in solution phase and co-crystallized onto a sample carrying target surface.
Since the matrix compound is used in excessive amounts, the resulting sample consists of matrix compound crystals with analyte molecules embedded into its crystal lattice. Utilization of matrix compounds increases ionization efficiencies dramatically, and also extends the area of applicability of these methods. Matrix-assisted laser desorption ionization (MALDI) [Karas, Hillenkamp, Analytical Chemistry, 60 (20): 2299-2301, 1988] is widely used for intact protein analysis and for protein identification based on the MS investigation of tryptic digests, besides polymer, nucleic acid and carbohydrate analysis. Main disadvantages of MALDI include the low ion yield, production of predominantly singly charged ions and the fact that natural surfaces can only be investigated after deposition of matrix compounds.
Need for desorption ionization methods working under atmospheric conditions has been raised recently. Advantages of atmospheric pressure desorption ionization method include: (1) Samples are not introduced into vacuum regime of mass spectrometer, which makes analytical procedure faster and more flexible, (2) since the sample does not enter vacuum, there is no need for the removal of volatile components, such as water, (3) arbitrary objects can be investigated/analyzed this way, (4) biological systems including living organisms can be investigated in an in-vivo and in-situ manner, which feature allows the application of these methods for in situ tissue identification. Desorption ionization methods utilizing collimated beam of atoms, ions, molecules, or molecular clusters cannot be used under atmospheric pressure conditions, since particles cannot be accelerated to suitable velocities at high pressure due to consecutive collisions with gas molecules. Same phenomenon is also responsible for the extreme divergence of particle beams at higher pressures, which also hinders the formation of practical analytical beams.
Among the above described methods, only laser desorption ionization can be implemented at atmospheric pressure without dramatic changes in instrumentation, since laser beams do not interact with air molecules under the conditions of ionization. Atmospheric pressure MALDI was developed by Laiko et al. (2000), Anal. Chem., 72, pages 652-657; however the technique did not gain popularity due to low ion yield which problem is further increased by the 99% ion loss in atmospheric interface, and workplace safety issues generally associated with the use of laser in open experimental setups.
The recently developed desorption electrospray ionization (DESI) [Takats et al, Science, 2004] is taxonomically/phenomenologically the atmospheric pressure version of MCI technique described above. Both methods employ multiply charged solvent droplets as analytical beam, however in the case of DESI droplets are produced by electrospray and accelerated by supersonic gas stream instead of electrostatic field gradient. Nevertheless, DESI has fulfilled all expectations associated with atmospheric pressure desorption ionization methods, so it opened the door to the mass spectrometric analysis of arbitrary objects with regard to chemical composition, size and geometry. In the course of the DESI process, high velocity electrosprayed droplets impact with the sample surface. Impacting droplets dissolve molecules present on the surface, and emit secondary droplets which are charged. Charged secondary droplets carrying surface material produce ions finally following the well-known mechanisms of electrospray ionization.
Investigation of tissues by means of mass spectrometry has been pursued in two, fundamentally different ways. One approach was focused on the systematic characterization of compound groups present in tissues, while the other strategy concentrated on the fast, direct MS fingerprinting of tissues. Methods belonging to the first group generally start with homogenization and lysis of large amount of tissue, followed by selective extraction of compound group of interest (e.g. proteins or phospholipids, etc.). Compounds are separated by means of electrophoresis or chromatography, and then detected by mass spectrometry. Although these methods cannot be used for fast identification of tissues, they provide invaluable information on marker molecules characteristic to one or another type of tissue.
Fast mass spectrometric fingerprinting of tissues is generally achieved by desorption ionization methods described above. SIMS analysis of tissues gives characteristic spectra showing mainly phospholipid fragments, however, the technique works exclusively under high vacuum conditions, and hence it cannot be applied for in vivo analysis of tissues. MALDI analysis of tissue samples gives spectra featuring either ions of abundant proteins, or ions of common membrane lipids, depending on type of matrix compounds employed. Although both types of spectra are characteristic, and show unique features in case of malignant tumours, the method still cannot applied for in vivo analysis, since deposition of matrix compounds is incompatible with living organisms. Direct laser desorption ionization using infrared laser (Er-YAG or CO2) is a special case of MALDI, where water content of sample acts as matrix. This method is fully compatible with in vivo analysis (these infrared lasers are widely used in surgery), however tissue identification in this case has not been demonstrated until now. Recently developed DESI methods give spectra featuring various membrane lipids, which give characteristic patterns for a number of tissues. DESI analysis does not require any sample preparation, unlike MALDI, thus freshly cut surfaces of living tissue can be investigated. DESI analysis of living tissues, however, does not yield conclusive data, due to interference from blood and interstitial fluid leaking from surface being investigated. Further disadvantage of DESI analysis is the safety concern associated with the use of 4-5 kV DC in proximity of living organisms.
From the above analysis of the state of the art in MS it can be concluded that both abundant proteins and phospholipids give characteristic distribution in Dl mass spectra of various tissues, however for in vivo MS analysis there is no appropriate ionization method developed yet.
Ionization of condensed phase, non-volatile samples via rapid heating has been pursued since the late 1960's. Rationale of this effort was to employ sufficiently high heating rate to achieve disintegration rates comparable to rate of decomposition of analyte molecules. Friedman et al. have described successful ionization of amino acids and peptides by rapid heating in the early 1970's. Assumed mechanism of these experiments was associated with the direct disintegration of ionic species present in solid phase. Experimental implementation of these experiments was limited to contact heating of pure, crystalline analyte compounds. The search for more efficient methods of heating has led to application of lasers in mass spectrometric ion sources, and eventually to the development of various laser desorption ionization methods (including MALDI) described above.
Alternatively to laser heating, thermally assisted spray disintegration of solution phase compounds have also been studied (see for example Vestal et al, Anal. Chem. (1980), 52, pages 1636-1641. Since spray disintegration in a vacuum or inert atmosphere dramatically increases the rate of disintegration, intact molecular species were successfully transferred to gas phase this way. These methods were termed “thermospray” and were widely used in the late 1980's and early 1990's as HPLC-MS interfaces. Most thermal disintegration methods result in the formation of overwhelmingly neutral species; hence these methods were often combined with post ionization techniques. Post ionization has been traditionally carried out via electron impact (EI) or chemical ionization (CI). Recently, a similar approach has been introduced utilizing electrospray post-ionization of gaseous species obtained by laser ablation of samples (LAESI).
What is needed is an MS-based device, system and method which can be used for direct, in situ investigation of biological tissues, that does not harm organisms being investigated and gives mass spectra characteristics to different types of tissues in a relatively short timeframe, and that also can be used in the operating room, advantageously as an integrated part of one or more surgical tools or dissecting tools.